Re: [ROOT] Fitting Question

From: Eddy Offermann (eddy@rentec.com)
Date: Tue Oct 31 2000 - 23:33:29 MET


Ray,

Another (more general) way is to implemement these types of constraints
through Lagrangian multipliers. In your case minimize

f(x,p0,p1,p2,p3,p4) + mu * (p0+p4-C) where you have now an additional

variable mu to be minimized. A possible adavantage over parameter
substitution is that you will get the error matrix for the p's taking
the constraint into account. 

Eddy
> 
> Hi Ray,
> Fitting with constraints is not implemented.
> Reduce the number of parameters from 5 to 4 and in your function
> define an internal variable p4=C-p0
> 
> Rene Brun
> 
> 
> On Tue, 31 Oct 2000, Ray Fliller III wrote:
> 
> > 
> > Hello Rooters,
> > 
> >    Is it posible to fit a function and impose an equation of constraint 
> > between the variables, for example:
> > 
> >   I have a function f(x,p0,p1,p2,p3,p4) where the p's are my fitting 
> > parameters.  Is there a way to impose on the fit the constraint that 
> > p0+p4=C where C is some constant??  Or some other such constraint??
> > 
> > Thanks.
> > 
> > 
==============================================================================
> > Ray Fliller: rfliller@bnl.gov      Office Phone: (631)-344-6124 
> > C-A Accelerator Physics     
> > Building 911
> > Brookhaven National Lab
> > Upton, NY 11973                    Office: Room 211
> > 
==============================================================================
> > 
> 

Eddy A.J.M. Offermann
Renaissance Technologies Corp.
Route 25A, East Setauket NY 11733
e-mail: eddy@rentec.com
http://www.rentec.com



This archive was generated by hypermail 2b29 : Tue Jan 02 2001 - 11:50:36 MET